Abstract
As an effective means of acquiring knowledge resources, knowledge search is an important driving force for enterprises to obtain competitive advantages. Based on ambidextrous theory, organizational search theory and knowledge base theory, this study constructs a theoretical framework for the effect of ambidextrous knowledge search on technological innovation performance. And empirically tests the mediating mechanism of knowledge base with the data of knowledge-intensive enterprises. The results show that both exploratory and exploitative search have positive effects on technological innovation performance, and significantly, the positive impact of exploratory searches was higher than that of exploitative searches. Moreover, the balance and combination of knowledge search have a significant positive impact on technological innovation performance respectively. However, the effect of combination is stronger when two search dimensions are considered at the same time. There is a mediating effect between ambidextrous knowledge search and technological innovation performance. Specifically speaking, knowledge base plays a complete mediating role for ambidextrous balance whereas a partial mediating role for ambidextrous combination. The conclusions reveal the mechanism of ambidextrous knowledge search, which can provide a certain theoretical guidance for enterprises to search external knowledge and improve innovation practice.
Plain Language Summary
Given the synergistic effects as well as the potential negative effects of ambidextrous search, the key is to identify the mechanism between ambidextrous search and technological innovation. In this study, we have two objectives. First, we distinguish the two types of knowledge search by using patent data that span technological boundaries, and propose separate and interaction effects of exploitation and exploration search in the perspective of ambidexterity on technological innovation. Second, we develop and investigate the mediator role of knowledge base, through which firms can transfer the fresh knowledge to innovation benefits. To achieve the goals, we test our theoretical model using data from 231 Chinese listed firms. And find that the relationship between ambidextrous search and technological innovation performance is positively mediated by knowledge base. Particularly, the firm characterized by abundant knowledge base is the best at gaining the innovation benefits of an ambidextrous search strategy. Overall, this study enriches research as well as practice by providing insights into the effectiveness of ambidextrous search.
Keywords
Introduction
In the era of knowledge economy, the knowledge required for technological innovation is becoming more complex. It is difficult for enterprises to meet the diversified market needs by relying on internal R&D only. More seriously, the risks and challenges brought by the globalization of innovation and the reorganization of the international economic order give more pressure on firms. The continuous acquisition, integration and iteration of knowledge have become the key to enterprise technological innovation. Knowledge search, which refers to the firm’s effort to search for new knowledge from outside (Rosenkopf & Nerkar, 2001), has also become the main strategic choice for enterprises to solve the contradiction between innovation needs and the limited internal resource capabilities (Angelidou et al., 2022; Bedford et al., 2022). It is considered as the third way to improve the competitive advantage of organization except internal R&D and external acquisitions (Laursen, 2012). In order to improve the effectiveness of knowledge acquisition and innovation, enterprises need to weigh the potential of knowledge integration within organization and the cost of knowledge collaboration (Grigoriou & Rothaermel, 2017). However, exploration and exploitation (March, 1991) represent the highest demand of innovation and stimulate the generation of multi-layer nested paradoxes within organizations (Sheep et al., 2017). This “paradox” phenomenon always occurs throughout the process of knowledge search and innovation. Inevitably, enterprises face many competitive knowledge search activities such as width and depth (Laursen & Salter, 2006), market and technology (Sidhu et al., 2007), interactive and non-interactive (Roper et al., 2017), and local and boundary-spanning (Gao et al., 2020). Due to the scarcity of innovation resources and limited attention, different types of knowledge search activities are independent and contradictory (Andriopoulos & Lewis, 2009). For example, the exploration of fresh knowledge can continuously update the knowledge base but result in high search costs, leading enterprises fall into sub-optimal equilibrium. The exploitation of old knowledge can make firms more professional in existing technology field but cause path dependence. Therefore, the paradox implicit in different types of knowledge search methods is still a basic practical problem that enterprises urgently need to solve.
Specifically speaking, the innovation paradox is defined as the coexistence of contradictory elements and their unity of opposites embedded in the innovation process (Miron-spektor et al., 2011; Schad et al., 2019; Smith et al., 2017). The tension, which reflects the result of conflict between paradox elements, is the core of innovation paradox (Smith & Lewis, 2011). The paradoxical relationship between exploration and exploitation is mainly reflected in two aspects, namely process and result. The former one is reflected as a tension between proposal and implementation of creative ideas, while the latter between novelty and practicality of innovation results (Rosing et al., 2011). No matter which perspective, the innovation paradox between exploration and exploitation has a double-edged sword property. It brings uncertainty and ambiguity that may lead to anxiety, intensify internal conflicts, and impede organizational change. Yet effective paradox management can help enterprises achieve synergistic effect, improve organizational creativity, and realize sustainable development (Schad et al., 2016).
Ambidexterity is regarded as an effective measure to solve the innovation paradox in the field of organizational management. It is crucial for enterprises to maintain a sustainable competitive advantage that managing and coordinating the paradoxical activities. This is also the core issue of ambidextrous research (Tarba et al., 2020). The ambidexterity theory is an effective method to resolve the paradoxical contradiction between exploration and exploitation (Gibson & Birkinshaw, 2004). Specifically, ambidexterity is considered to be a management ability to face the tension of exploration and exploitation, which can obtain organizational flexibility to adapt to an uncertain environment through the coexistence and balance at a high level (Simsek, 2009). With the popularity of the open innovation paradigm, boundary-spanning search provide a possibility for enterprise to utilize internal and external resources and knowledge at the same time. To some extent, this search activity can breakthrough internal resource constrains and brings fresh perspective for ambidextrous research. In the context of open innovation, the connotation of ambidexterity extends from the inside to the outside of the organization. Hence, it is not only necessary to consider the “balance” and “combination” of exploration and exploitation, but also need to probe their coexistence inside and outside the organization (O’Reilly & Tushman, 2013; Stettner & Lavie, 2014).
The existing literature has generally recognized the strategic significance of knowledge search for innovation performance (Katila & Ahuja, 2002; Laursen & Salter, 2006; Ye et al., 2023), especially under the circumstance of open innovation (Chesbrough & Bogers, 2014; Gao et al., 2020; Saura et al., 2023; D. Zhang et al., 2017). However, there are few studies based on ambidextrous perspective (D. Zhang et al., 2017). On the one hand, the ambidextrous research of knowledge search is still in its infancy stage. Existing research lacks a systematic understanding of the innovation mechanism of ambidextrous knowledge search. On the other hand, the open innovation context makes competitive knowledge search activities can coexist in multiple latitudes (Cantarello et al., 2012), which increases the complexity of ambidextrous research. Besides, ambidextrous search also suffers from resource constraints and coordination problems with cost and risk (Sidhu et al., 2007). Therefore, based on the theories of organizational search, ambidexterity, and knowledge base, this article analyzes the differential impact of distinct knowledge search strategies on ambidextrous innovation under the premise of resource constraints. And then explores the influence mechanism of knowledge base on the relationship between the two from the perspective of ambidexterity.
Given the synergistic effects as well as the potential negative effects of ambidextrous search, the key is to identify the mechanism between ambidextrous search and technological innovation. In this study, we have two objectives. First, we distinguish the two types of knowledge search by using patent data that span technological boundaries, and propose separate and interaction effects of exploitation and exploration search in the perspective of ambidexterity on technological innovation. Second, we develop and investigate the mediator role of knowledge base, through which firms can transfer the fresh knowledge to innovation benefits. To achieve the goals, we test our theoretical model using data from 231 Chinese listed firms. And find that the relationship between ambidextrous search and technological innovation performance is positively mediated by knowledge base. Particularly, the firm characterized by abundant knowledge base is the best at gaining the innovation benefits of an ambidextrous search strategy. Overall, this study enriches research as well as practice by providing insights into the effectiveness of ambidextrous search.
The remainder of the paper is organized as follows. In the next section, we review the relevant literature of our study and provide the development of hypotheses based on this theoretical background. In the third section, we describe our empirical approach and outline our sample of firms in the high-tech industry, followed by the description of our results in the fourth section. Following this, the results are discussed and potential implication and limitation of this study are outlined in the fifth section. And finally, the paper makes a brief conclusion in the last section.
Literature Review and Hypotheses Development
Knowledge Search Strategy
Search strategy, according to behavioral research and innovation literature, is considered as a problem-solving activity in which firms develop innovations through identifying and integrating novel knowledge (Knudsen & Levinthal, 2007; Angelidou et al., 2022). The knowledge-based view believes that knowledge resources are the foundation and source of technological innovation. And the behavioral concepts of exploitation and exploration offer important insights to understanding the scope of knowledge search (Ng et al., 2019). Hence, it is widely recognized in innovation studies that the firm’s innovative performance is a function of the knowledge targeted and accessed by the search strategies (Ehls et al., 2020; Jung & Lee, 2016; Ko et al., 2021). With the increasing complexity of innovation, it is less and less possible for companies to achieve innovation by only relying on their own knowledge. More enterprises embrace the open innovation to search heterogeneous knowledge for discovering new technologies, developing new products, or creating new business processes (Chesbrough, 2003; Huizingh, 2011). Existing research has confirmed the role of knowledge search in promoting technological innovation (Antonelli & Colombelli, 2015; Saura et al., 2023).
On the one hand, exploitative search has the advantages of low cost and high reliability that can seek solutions from past experiences and conventions in familiar knowledge areas. It can develop new ways of applying existing knowledge to maximize the benefits of limited resources by reusing existing technologies and knowledge. Moreover, exploitative search also form a self-reinforcing and accumulative effect through repeated practice and gradual inheritance, which can improve the sensitivity and judgment ability of enterprise and make organization more professional in the existing fields. On the other hand, the advantage of explorative search is embodied in the acquisition of heterogeneity and diversity knowledge (Hume & Hume, 2015), which can break through path dependence and experience constraints to a certain extent (C. Kim & Park, 2013). It is deemed as an indispensable organizational learning process that help enterprises survive and develop in a rapidly changing external environment with limited internal resources. Enterprises can break existing knowledge pattern to form a new framework and achieve technological leap by acquiring more quantitative and higher qualitative heterogeneous knowledge (Classen et al., 2012). Various knowledge search channels bring more innovation opportunities to enterprise. For example, knowledge obtained from stakeholders in the innovation chain such as consumers or suppliers can provide fresh ideas and sources for innovative activities (Love et al., 2014). Hence, to some extent, both exploration and exploitation search strategies have positive effects on technological innovation.
Ambidextrous Knowledge Search and Technological Innovation
Yet, while exploration and exploitation search offer a potential for technological innovation, we must recognize the challenges of acquiring and integrating knowledge across organizational and technological boundaries (Enkel & Gassmann, 2010). Due to limited attention and scarcity of resources, two types of search activities need to compete with each other within organization (Lavie & Rosenkopf, 2006; Yan et al., 2016). That is to say, search activities are not costless and require substantive resource commitments and attention (Y. Zhang & Li, 2010). Moreover, the likelihood of “over-search” effect rises if a company performs one search activity only.
As mentioned above, ambidextrous theory can resolve this paradox. Specially speaking, organizational ambidexterity is defined as “the ability of an organization to both explore and exploit” (O’Reilly & Tushman, 2013). Researchers find that enterprise with ambidextrous ability can survive in the long term (García-Lillo et al., 2016). Nevertheless, how to achieve ambidexterity? The scholars discussed mainly focus on two dimensions, namely balance and combination (Cao et al., 2009). The balance dimension focus on the idea of decomposing conflicts to avoid competing resources by dynamically changing explorative and exploitative activities at different stages. Whereas the combination dimension puts emphasis on the idea of accepting conflicts that the two activities can support each other in complementary fields (Raisch et al., 2009). In a whole, ambidextrous knowledge search makes up the shortcomings of a certain activity.
In the perspective of balance dimension, enterprises can alter knowledge search methods according to different stages of innovation activities so as to achieve a balance of knowledge search in various time or space. If a firm pays too much attention to a certain knowledge search dimension, it is prone to extreme situations. For example, the exploration of fresh knowledge can continuously update the knowledge base but result in high search costs and vicious circle, leading enterprises fall into sub-optimal equilibrium and capability trap (Andriopoulos & Lewis, 2009). Moreover, the exploitation of old knowledge can make firms more professional in existing technology field but cause path dependence and core rigidity that make enterprises unable to quickly adapt to market environmental changes. Hence, the balance of exploration and exploitation on the relative scale can avoid the occurrence of the above-mentioned extreme situations and help enterprise improve the technological innovation performance. Thus, we predict that:
H1a: There is a significant positive correlation between the balance dimension of ambidextrous knowledge search and technological innovation.
As far as the combination dimension is concerned, there is no need to compete for innovative resources if exploration and exploitation occur in complementary fields. They produce a joint effect by supporting each other in the search process (Gupta et al., 2006). Specially speaking, enterprises have a deeper understanding of their own knowledge base and more accurate positioning in the industry by repeating use of existing knowledge and resources. This not only improves the efficiency of exploring novel knowledge, but also helps organizations to grasp the market trend and search direction. On the contrary, the homogeneous knowledge searched by exploration activity contributes to break the existing knowledge pattern and form a new framework. It increases the combined opportunity of knowledge elements as well as enhances the adaptability of enterprises to dynamic environment by expanding knowledge base. Hence, we hold that exploration and exploitation activities can often complement each other and enhanced firm performance. On the basis of these arguments, we predict that:
H1b: There is a significant positive correlation between the combined dimension of ambidextrous knowledge search and technological innovation.
Ambidextrous Knowledge Search, Knowledge Base, and Technological Innovation
Knowledge-based theory deems knowledge as the key resource for enterprise to survive (Yao et al., 2021). While ambidextrous knowledge search provides more fresh knowledge, it also puts forward new requirements for enterprises to choose appropriate knowledge and reorganize existing knowledge organization (Rui & Lyytinen, 2019). Here, the current status of knowledge base of a company is a determining factor in the extent to which the firm can understand and effectively encode the acquired knowledge (Shafique, 2013). Previous studies have shown that the knowledge base affects the firm’s technological innovation (Zhou & Li, 2012), but have failed to show to what extent the company’s existing knowledge base mediates the influence of ambidextrous knowledge search on technological innovation. The technological innovation is a knowledge integrated process involving multi-disciplinary fields, which requires complementary and specialized knowledge (Madhavan & Grover, 1998). However, as the knowledge resources required for innovation become more complex, knowledge gap is often formed within enterprise. One of the reasons is that internal resources are scarce both in quantity and quality, and the other is that the tacit knowledge resources are underutilized. Therefore, enterprises have to expand their knowledge base by searching heterogeneous knowledge outside the organization or conducting R&D activities within the organization to reuse existing knowledge effectively (Moorthy & Polley, 2010).
To achieve technological innovation performance, enterprises need to implement correct knowledge search strategy so as to obtain novel knowledge resources. Meanwhile, the matching internal knowledge base is also necessary to promote the integration of innovation resources (Rui & Lyytinen, 2019). Based on the coverage of knowledge and the familiarity with knowledge, extant research divides the knowledge base into two dimensions, breadth and depth (Zhou & Li, 2012; Yao et al., 2021). The former usually refers to the diversity and heterogeneity of the knowledge base, while the latter represents the complexity and expertise of the knowledge structure (Mannucci & Yong, 2018). The choice of knowledge search strategy needs to rely on the knowledge base. At the same time, the search strategy will also promote the improvement of the knowledge breadth and depth.
There is no doubt that the knowledge search activities, no matter which dimension, will has a positive impact on the knowledge base to a certain extent. Specifically, explorative search mainly focuses on the heterogeneous knowledge beyond the current technology domain or organizational boundaries that can broaden the breadth of knowledge base. While exploitative search places more emphasis on similar knowledge that can promote a deeper understanding of some specific knowledge and deepen the depth of knowledge base. However, the focus of these activities is different and the impact of their balance and combination on the knowledge base will also be different. Therefore, we propose the following hypotheses regarding the impact of ambidextrous knowledge search on knowledge base.
H2a: Both balance and combined knowledge search have significant positive effects on knowledge breadth, but combined search has a stronger effect.
H2b: Both balance and combined knowledge search have significant positive effects on knowledge depth, but balance search has a stronger effect.
The view of innovation reconstruction believes that innovation is a process in which an enterprise continuously reorganizes existing knowledge elements. In the process of knowledge search and innovation, some influencing factors such as the quality of searched knowledge, the speed of knowledge absorption and the efficiency of knowledge integration, will lead to different innovative outputs (Appio et al., 2017; M. Wang & Wang, 2022). In the perspective of knowledge base, the broader a business has, the higher the likelihood of combining different knowledge elements and the greater the ability to develop new technologies. Firstly, the combined potential between knowledge elements will be exhausted as it is reused. Therefore, the high breadth of knowledge base could provide more paths to explore different problem solutions and new frontiers (D Este, 2005). Secondly, there is knowledge spillover between various technical fields. A diversified knowledge base can improve innovative efficiency and increase innovation output, so as to gain competitive advantage in a rapidly changing environment (J. Zhang & Baden-Fuller, 2010). Thirdly, a broad knowledge base can consolidate the absorptive capacity that is conducive to acquiring, understanding, and learning novel knowledge, thereby improving the performance of technological innovation.
The knowledge breadth mastered by enterprise provides the necessary conditions for cross professional integration in innovation process (Yayavaram & Ahuja, 2008). First of all, abundant knowledge resources and spillovers in different fields can provide a rich creative pool to ensure the efficiency of transformation of knowledge resources into innovative achievements. In the second place, the breadth of knowledge base is also beneficial to carry out extensive learning and improve the sharing ability of tacit knowledge within organization. Last but not the least, diverse knowledge can stimulate firm’s foresight thinking model and improve the adaptability to deal with uncertainty.
Meanwhile, the depth of knowledge base is also necessary for technological innovation. As a long term and continuous process of knowledge accumulation, knowledge depth reflects the familiarity with existing knowledge elements in a particular technology field. In this process, the procedure of knowledge transformation, assimilation, and application is gradually standardized and systematized. On the one hand, the ability of enterprise to absorb and utilize new knowledge will be significantly enhanced with the improvement of knowledge depth (Xu et al., 2019). The routine and codified knowledge developed during organizational learning form the basis for technological innovation (Menguc et al., 2014). The technological innovation arises from the novel knowledge or the combination of the existing and new knowledge. Therefore, competitive new technologies and products can only be developed by continuously deepening its knowledge in a specific field and improving its core technology level. On the other hand, knowledge depth promotes the use of existing fragmented and isolated knowledge within organization (Wei et al., 2022). It helps to form an integrated and systematic knowledge system that enhancing the flexibility and creativity of enterprise. In addition, the knowledge depth can also improve the ability of identifying novel knowledge elements that enhancing the efficiency of knowledge absorption and integration. Thus, we presume that the knowledge base of enterprise has a positive impact on technological innovation performance.
H3a: Knowledge breadth is positively related to technological innovation.
H3b: Knowledge depth is positively related to technological innovation.
As mentioned above, the ambidextrous knowledge search activities promote the changes in existing knowledge base, which can not only broaden knowledge breadth but also enhance knowledge depth. Thus, the evolving knowledge base can provide abundant knowledge sources for technological innovation. However, knowledge search requires costs and large amount of irrelevant knowledge also causes problems such as redundancy (Flor et al., 2018). For example, scarcity of innovative resources and limited attention of managers believes that excessive diversified knowledge hinder the enterprise from forming the core technologies because large amount of redundant knowledge makes it difficult to communicate, coordinate, and integrate within organization. Besides, there is inherent risk in the development of unrelated technology (Ferreras-Méndez et al., 2015). Furthermore, the marginal benefit of R&D investment will decrease as the key technology is developed thoroughly when the knowledge search is limited to a specific technical field.
These problems caused by knowledge search can be alleviated by knowledge integration through the use of organizational capabilities. Enterprises obtain large amount of knowledge and information by using knowledge search strategy. However, as a static resource, these will not be automatically transformed into innovation performance but need to go through a process of identification, transformation, and configuration within organization. It is well known that the essence of innovation is a process of searching, absorbing, transforming, and reorganizing fresh knowledge and finally outputting products and services (Carnabuci & Operti, 2013; Gao et al., 2022). Nevertheless, the knowledge acquired through knowledge search is often disordered and scattered. It is difficult to share and reconstruct within the organization without transformation process, in which the knowledge base of enterprise plays a critical role. That is to say, the resources acquired by knowledge search activities can be transformed into the firm’s inner knowledge base through learning, absorbing, and integrating, and then serving technological innovation.
To sum up, the knowledge search activities promote the knowledge base that directly affects the technological innovation. Therefore, we argue that the knowledge base plays a mediating role between ambidextrous knowledge search and technological innovation and propose the following series of hypotheses.
H4a: Knowledge breadth mediates the relationship between balance knowledge search and technological innovation.
H4b: Knowledge breadth mediates the relationship between combined knowledge search and technological innovation.
H4c: Knowledge depth mediates the relationship between balance knowledge search and technological innovation.
H4d: Knowledge depth mediates the relationship between combined knowledge search and technological innovation.
Next to the proposed relationships between ambidextrous knowledge search behavior, knowledge base, and technological innovation performance, earlier research suggests that organizational and sociodemographic factors also additionally influence this relationship (e.g., Schnellbacher & Heidenreich, 2020; W. Zhang et al., 2019). Hence, the study includes firm size, firm type, gender, and managerial experience as controls. The inclusion should strengthen the robustness of the hypotheses tests. In Figure 1 the research model’s framework is summarized.
The research framework.
Methodology
Sample and Data Selection
The hypotheses were tested using the data of listed companies from China in the industry of computer, communication, and electronic equipment manufacturing. The selection of this context is mainly based on the following considerations. First, this high-tech industry is characterized by knowledge intensive, technology dependence, and high R&D investment. Compared with traditional industries, the sample in this rapidly developed industry is sufficient. Second, the data of listed companies is objective and standardized that make the research reproducible. Third, we select the data from different database to reduce homology bias. For example, the patent data is extracted from State Intellectual Property Office (SIPO) and Derwent Innovations Index (DII), whereas financial and disclosure data is derived from CSMAR. After removing observations with missing variables, *ST and ST (special treatment of the listed companies with delisting risk warning like and other risk warning because of abnormal financial status or other conditions), and negative assets, the useful sample of 231 listed companies is obtained.
Measurement
Dependent Variable
Technological Innovation Performance (TIP)
There is no unified opinion in academic on the measurement of innovation performance, but mainly focus on two methods of patent and questionnaire. The patent data, as the most direct and quantifiable achievements of technological innovation, is becoming an important data source for scholars to conduct innovation research because its objective review mechanism, uniform standard, and strong availability. Previous studies (Messeni Petruzzelli et al., 2015; Sampson, 2007) have demonstrated that patent citation is a good indicator for estimating firms’ innovation performance. However, patent citation data are not available in the SIPO, we could not evaluate the effects of patent citations. Besides, the number of patent citations rises over time and may be influenced by short technology life cycles (Haupt et al., 2007). In light of these issues, this study draws on previous research (Hall et al., 2005; C. Wang et al., 2020) and adopts the annual number of invention patent applications as a proxy indicator, which has been proven to be suitable for explaining technological innovation performance.
Independent Variables
Ambidextrous Knowledge Search
Studies have shown that there is a high correlation between patent citations and actual knowledge flows (Singh & Agrawal, 2011). Hence, we intend to use patent citations measuring the variables of knowledge search activities. Yet the lack of citation information in Chinese patent data makes it impossible. So, we use the international patent classification code of applied patent as a proxy variable. Specifically, this study focuses on the occurrence of knowledge search behaviors and divides it into exploration search and exploitation search based on the degree of technical boundaries. Referring to the measurement of existing research (Lodh & Battaggion, 2015), exploration search is measured by the number of new IPC codes, that appear in the firm’s annual patent applications, while exploitation search is evaluated by the number of old IPC codes. Furthermore, the measure of ambidextrous knowledge search is based on above two kinds of search activities (Cao et al., 2009). The balance dimension (BD) is expressed as the absolute value of the difference and the combined dimension (CD) is represented by the sum of the two. In other words, the former is calculated by |exploration − exploitation| while the latter is expressed by (exploration × exploitation).
Knowledge Base
Drawing on previous studies about technical knowledge diversity (M. Kim et al., 2016; Lodh & Battaggion, 2015; Moorthy & Polley, 2010), this study uses four-digit IPC codes as the basis to calculate firm’s knowledge base, measured by a method of modified HHI (Herfindahl-Hirschman Index). Assuming that all patents of a company is distributed in n categories, and p is the proportion of patents in i category to total patents. The calculation formula is as follows:
Among the formula, TKb and TKd represent the breadth and depth of knowledge base respectively. Different from the measurement of knowledge search, we use the data of patents that enterprise owned rather than applied over the last three years to reduce the bias caused by the changes of technological environment.
Control Variables
In line with previous studies, we choose firm size, managerial experience, top manger gender, and firm type that may have an impact on technological innovation performance as control variables. Specifically, we controlled for firm size by using the log value of total assets as larger firms have more resources (Aliasghar et al., 2019). The managerial experience (ME) is defined as the average age of top managers, since elder managers may have more experience that might affect the selection of search strategy. Moreover, we also controlled for top manger gender (TMG), using a binary variable which is assigned 1 if there are women in top manager team. Meanwhile, we differentiate firm type because that state owned firms may have more support policies. It is a binary variable that assigned as the value of 1 if state owned, and 0 otherwise. Table 1 shows a summary of the basic descriptive statistics
Descriptive Statistics of the Variables.
Results
As shown in Table 2, we firstly report the correlations. The table shows the presence of the positive correlation coefficient between any two main variables, which provide a rough support for research hypotheses. Moreover, the variance inflation factor (VIF) values are used to test the possibility of multicollinearity. There is no serious multicollinearity problem with the evidence of the largest value of VIF (3.2). It is generally accepted that the likelihood of multicollinearity problem is minimal when the VIF values are well below the recommended ceiling of 10 (Kang & Kang, 2014).
Correlations of Explanatory Variables.
p < .10. **p < .05. ***p < .01.
The Influence of Ambidextrous Knowledge Search on Technological Innovation
We employed ordinary least squares regression analysis in Stata to test the hypotheses proposed in this study. And the hierarchical moderated method was adopted to observe the variation trend of the variable coefficient. The regression results of the main effect are shown in Table 3.
Regression Results of the Direct Effect.
p < .10. **p < .05. ***p < .01.
The benchmark model is Model 1 with all control variables. Its result depicts the influence of control variables on technological innovation performance. The result shows the significant positive effect of firm size and negative impact of management experience. Specifically, the firm size has a positive effect on technological innovation, in line with the Schumpeter’s hypothesis, because large enterprises have more resources (Aliasghar et al., 2019). However, the management experience has a negative impact on technological innovation performance because that the experience in a certain field easily leads to habitual thoughts hampering innovation.
The results revealed that a certain type of knowledge search, whether exploitation or exploration, has a positive impact on technological innovation performance. To be specific, the regression coefficients of exploitation and exploration search in Model 2 and Model 3 show their positive effect on technological innovation (β = 3.424, p < .1; β = 4.470, p < .01). Furthermore, Model 4 simultaneously takes two knowledge search types into consideration, the test result not only confirms the above hypotheses again, but also shows a greater effect of exploration search. We use the T test method to test the significance of this different influence between exploration and exploitation search. The results show that the standardized regression coefficients of exploitation and exploration search are 0.121 (t = 1.79, p < .1) and 0.175 (t = 2.82, p < .01), respectively, indicating the significantly stronger positive effect of exploration search. That is to say, it is important for knowledge intensive enterprises that heterogeneous knowledge in various fields collides to generate new sparks, forms fresh ideas, and then promotes innovation performance.
From the perspective of ambidexterity, the regression result of Model 5 shows that the BD (balance dimension) has a negative effect (β = −6.016). This indicates that the imbalance between exploration and exploitation leads to the worse technological innovation performance. But this result is not significant, H1a is not supported. Meanwhile, the CD (combined dimension) has a markedly positive impact (β = 5.408, p < .01) on technological innovation. This proves that exploration and exploitation activities can produce a joint effect in the knowledge search process, which supports H1b.
The Mediating Role of Knowledge Base
In line with previous research (Gao et al., 2022), we use the step-by-step method to test the mediating effect of knowledge base. The test results of main effects have shown the significant relationship between ambidextrous knowledge search and technological innovation, satisfying the first step of mediating effect test. In the next, we need to examine the relations between dependent and mediating variables, mediating and independent variables, as well as the change after adding mediators.
As depicted in Table 4, the results of Model 6 to 11 demonstrate the positively effect of ambidextrous knowledge search on the firm’s technological knowledge base. To be specific, the regression coefficients in Model 6 and Model 7 prove that a single dimension of ambidextrous knowledge search has positively impact on the breadth of knowledge base, whether it concerns balance dimension (β = .125, p < .01), or combined dimension (β = .060, p < .01). Likewise, the results of Model 9 (β = .067, p < .01) and Model 10 (β = .023, p < .01) show similar evidence about the effect of ambidextrous knowledge search on the depth of knowledge base. In brief, there is a significant and positive effect of ambidextrous knowledge search on knowledge base when it only comes to one dimension of ambidexterity. These outcomes satisfy the second precondition of stepwise test. Moreover, the regression results of Model 8 and Model 11 certify that the influences of two dimensions of ambidextrous knowledge search on knowledge base are significantly different. However, no matter for the breadth or the depth of knowledge base, the effect of CD is always stronger than that of BD. Hence, H2a is supported, but H2b is only partially supported.
Regression Results of the Mediation Effect.
p < .10. **p < .05. ***p < .01.
Additionally, the coefficients in Model 12 and Model 13 certify the significant role of the knowledge base breadth and depth to enhance the technologically innovative performance (β = 40.112, p < .05; β = 61.664, p < .01), supporting H3a and H3b. These results meet the third precondition of stepwise test and the mediating effect is proved by Model 14 to 17. Specifically, the coefficients of TKb and TKd in Model 14 and Model 15 are positive and significant (β = 62.884, p < .01; β = 92.158, p < .01) whereas the coefficient of BD is insignificant. Therefore, the knowledge base, whether breadth or depth, has a mediating effect between the balance dimension of ambidextrous knowledge search and technological innovation performance, which support H4a and H4c. There is a similar result in Model 16 and Model 17. The significance of CD becomes weaker after adding the mediator variables, which support H4b and H4d. Moreover, knowledge base plays an absolute mediation role between BD and technological innovation but a partial mediation role between CD and innovation performance. Furthermore, the mediating effect of TKb accounts for about 83.02% of the total effect while TKd accounts for 44.83%.
Discussion
Theoretical Contributions
Through the lens of ambidexterity, long-term success depends on the ability to explore new opportunities and to exploit existing capabilities (Andriopoulos & Lewis, 2010; Schnellbacher and Heidenreich, 2020). Firms work under pressures related to innovation because the lack of resources and the intensive context (Aliasghar et al., 2019; Brem et al., 2016). Hence, it is essential for organizations to creating fresh ideas by searching internal and external knowledge (Paruchuri & Awate, 2016; Ehls et al., 2020). The purpose of this research is to explore the effect of ambidextrous search on technological innovation and how the knowledge base shapes this relationship. In this study, we adopted a hierarchical regression method to understanding ambidextrous knowledge search in organizations. Based on the theories of ambidexterity, knowledge base and organizational search, this research proposed several specific hypotheses that hope to generate new insights into the mechanisms. With the findings about the different effect of ambidextrous search on technological innovation as well as the mediating effect of knowledge base, this paper provides several primary contributions that bear significant implications for research concerning knowledge search strategies.
First, this study contributes to knowledge search by examining the effect of different dimensions of knowledge search in the perspective of ambidexterity. Research on the search strategy has largely focused on the dimensions of breadth and depth (Katila & Ahuja 2002; Laursen & Salter 2006), while recently has begun to study the ambidextrous search based on exploratory and exploitative activities (Cao et al., 2009). In line with previous empirical research (Y. Zhang et al., 2021), the result shows that the knowledge search behaviors (whether exploration search or exploitation search) have a directly significant positive impact on technological innovation. Beyond the individual effect of exploration search and exploitation search, this study creatively examines the effect of ambidextrous search in term of the balanced and combined dimensions on technological innovation. Our results prove the active effect of combined dimension of ambidextrous knowledge search but that effect of balance dimension is not confirmed. This finding indicates that in the high-tech industry firms can gain the technological innovation benefits through interacting exploration search and exploitation search rather than balancing them. That is to say, exploration and exploitation search activities may take place without necessarily compete for the same resources, which can often complement each other and lead to enhanced technological innovation performance. Hence, this study pushes the research frontiers on how firms gain the technological innovation benefits from ambidextrous knowledge search.
Second, this paper contributes to a better understanding of the channels through which knowledge search impact on technological innovation. As consistent with our expectations, knowledge base does mediate this relationship. That is to say, the knowledge searched from external partners needs to be transferred into the internal knowledge base and then generate innovative performance. This research divides the knowledge base into two dimensions, breadth and depth, and test their mediating roles respectively. The outcome says that knowledge search have a supporting role in technological innovation, enterprises need to transform the acquired knowledge into the internal knowledge base through assimilation and absorption for better technologically innovative performance improvement. In a whole, the positive impact of ambidextrous knowledge search on technological innovation can be achieved by building and improving the knowledge base. Additionally, the findings related to the knowledge base are in line with and enrich the literature on knowledge accumulation and innovation performance (e.g., Lee & Huang 2012). The study accentuates that such knowledge base can be fueled by rather explorative as well as exploitative knowledge search activities, following an idea already explored by Katila and Ahuja in the year 2002.
Last but not least, this study sheds some lights on which kind of search type can better improve the technological innovation performance. On the one hand, there exist differences when exploration or exploitation search affect technological innovation. Namely, exploration search has a significantly positive greater impact on technological innovation than that of exploitation search. On the other hand, the importance of two dimensions of ambidexterity is different in technological innovation. The combined dimension plays a vital role whereas the impact of balance dimension is insignificant. So does its effect on knowledge base. Overall, this study finds that although both exploration search and exploitation search contribute to technological innovation, firms cannot easily gain the innovation benefits of an ambidextrous search strategy. The findings significantly improve our understanding on the implications of search pattern to technological innovation.
Practical Implications
Our results provide some insights for managerial practice on how ambidextrous knowledge search affect technological innovation performance and highlights the important role of knowledge base between them. To some extent, the research conclusions can provide ideas for knowledge intensive enterprises in practice, such as carrying out knowledge search activities, improving internal knowledge base, and transforming external resources into innovative outcomes.
On the one hand, our research clarifies the knowledge search and knowledge base in the perspective of ambidexterity. Based on exploration and exploitation search, the ambidextrous search is divided into balance and combined dimension. And we further discover the difference of the effect between the two types and two dimensions of knowledge search on technological innovation, enriching the understanding of innovative search. Managers should be aware of the important strategic role of knowledge search in improving technological innovation performance and appropriately focus on the combination of two types of knowledge search. Thus, to facilitate technological innovation a firm should actively commit substantive resources to support these search activities. Moreover, enterprises should pay attention to the high combined levels of exploration and exploitation that can enhance innovation performance through leveraging the complementary resources. To some extent, increasing the search level of novel knowledge through external sources can promote the performance of technological innovation. For example, the ways such as establishing R&D centers or strategic alliances in cooperation with stakeholders, strengthening the sense of participation of consumers can be widely used to obtain heterogeneous knowledge and innovative resources.
On the other hand, knowledge base is fundamental for enterprise to carry out innovative activities. Therefore, in addition to paying attention to the direction and extent of knowledge search strategy, managers should also attach importance to the accumulation of knowledge base. Only there is a high degree of matching between them, the enterprise can efficiently build core capabilities, realize value co-creation, and form a competitive advantage. Most importantly, it is necessary to improve the knowledge base in the perspective of both breadth and depth according to firm’s technical characteristics and strategic needs. For instance, enterprise can enrich the existing knowledge base by introducing high-level talents, cooperating with research institutions, and establishing independent innovation platforms. The wide and deep knowledge base can provide the matched knowledge and ability for searched heterogeneous knowledge, which essential for technological innovation.
Limitations and Future Research Directions
Furthermore, our work also has some limitations that future research can address. First of all, this study used the patent data to measure main variables. This limited the research in a single industry to avoid the difference of patent levels in each industry. In addition, this paper tested its hypotheses by using cross-section data from the CSMAR and DII. A longitudinal study and a large sample in different industries may shed further light on the examination of these relationships. Moreover, it might be interesting for subsequent researches to address how other factors affect the interactive mechanism between ambidextrous knowledge search and technological innovation performance. This research only focused on the mediating role of knowledge base. The follow-up research could try to explore the influence of moderator factors on the main research framework. It is also possible to find different theoretical perspectives or methods to reveal research questions more deeply.
Conclusion
To sum up, knowledge search is an effective method to promote firm innovation performance (M. Wang & Wang, 2022). While exploration and exploitation knowledge search help firms gain the innovative benefits, such knowledge search pattern also suffers from some constraints such as limited resources, coordination problems, organizational slack, and formalization (W. Zhang et al., 2019). Hence, how can firms gain the benefits form ambidextrous knowledge search pattern is not only an urgent problem to be solved in the practice of enterprise innovation but also represents a critical research question in the domain of open innovation literature. The purpose of this research is to explore the effect of ambidextrous knowledge search on technological innovation performance and how a firm’s knowledge base shape this relationship. The empirical results of his study suggest that the current status of knowledge base plays an important role in realizing the synergistic effect of ambidextrous knowledge search on technological innovation performance. The findings enrich the understanding of ambidextrous knowledge search in theory, and also provide some guidance for enterprises designing optimal knowledge search strategy to facilitate technological innovation in practice.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The Social Science Foundation of Hebei Province (Grant No. HB22GL007).
Data Availability Statement
The data underlying this article were mainly accessed from CSMAR. All data included in this study are available upon request by contact with the corresponding author.